Cellular/Molecular

Transcriptional Control of KCNQ Channel Genes and theRegulation of Neuronal Excitability

Mariusz Mucha, Lezanne Ooi, John E. Linley, Pawel Mordaka, Carine Dalle, Brian Robertson, Nikita Gamper,and Ian C. WoodInstitute of Membrane and Systems Biology, Faculty of Biological Sciences, University of Leeds, Leeds, LS2 9JT, United Kingdom

Regulation of the resting membrane potential and the repolarization of neurons are important in regulating neuronal excitability. Thepotassium channel subunits Kv7.2 and Kv7.3 play a key role in stabilizing neuronal activity. Mutations in KCNQ2 and KCNQ3, the genesencoding Kv7.2 and Kv7.3, cause a neonatal form of epilepsy, and activators of these channels have been identified as novel antiepilepticsand analgesics. Despite the observations that regulation of these subunits has profound effects on neuronal function, almost nothing isknown about the mechanisms responsible for controlling appropriate expression levels. Here we identify two mechanisms responsiblefor regulating KCNQ2 and KCNQ3 mRNA levels. We show that the transcription factor Sp1 activates expression of both KCNQ2 andKCNQ3, whereas the transcriptional repressor REST (repressor element 1-silencing transcription factor) represses expression of both ofthese genes. Furthermore, we show that transcriptional regulation of KCNQ genes is mirrored by the correlated changes in M-currentdensity and excitability of native sensory neurons. We propose that these mechanisms are important in the control of excitability ofneurons and may have implications in seizure activity and pain.

IntroductionRegulation of the resting membrane potential and the repolariza-tion of neurons are important in regulating neuronal excitability.One ionic current that plays a key role in stabilizing neuronalactivity is the M-current (IM), a slowly deactivating, non-inactivating potassium current first identified nearly 30 years agoas the one underlying the excitatory effect of acetylcholine(Brown and Adams, 1980). The M-current is produced by theaction of Kv7 channels encoded by members of the KCNQ genefamily KCNQ1–KCNQ5, each of which encodes an individualpotassium channel subunit Kv7.1–Kv7.5 (Jentsch, 2000;Robbins, 2001).

Of the five KCNQ gene family members, KCNQ2 and KCNQ3are particularly important for regulating neuronal activity be-cause these subunits are expressed widely throughout the CNSwith expression patterns that almost entirely overlap (Tinel et al.,1998; Yang et al., 1998). Kv7.2 and Kv7.3 form functional hetero-multimers that are believed to represent a major M-channel iso-form in the CNS and peripheral nervous system (Delmas andBrown, 2005). The importance of the appropriate functioning of

Kv7.2 and Kv7.3 in the nervous system is highlighted by the factthat mutations in KCNQ2 and KCNQ3 are associated with be-nign familial neonatal convulsions, an autosomal dominant neo-natal epilepsy (Charlier et al., 1998; Singh et al., 1998; Yang et al.,1998; Biervert and Steinlein, 1999). The M-current is also oneof the key players in nociceptive transmission; inhibition ofM-current leads to membrane depolarization and hyperexcit-ability of nociceptive neurons (Passmore et al., 2003; Crozier etal., 2007; Linley et al., 2008; Liu et al., 2010), an important deter-minant of many pain conditions. Pharmacological openers ofKv7.2/Kv7.3 channels therefore now represent important analge-sic targets (Surti and Jan, 2005; Gribkoff, 2008; Wickenden andMcNaughton-Smith, 2009) because current therapies are not ef-ficacious for the majority of inflammatory and chronic painconditions.

Despite their importance, very little is known about the mech-anisms responsible for regulating KCNQ2 and KCNQ3 expres-sion. Here we use a functional assay to identify importantregulatory regions within the KCNQ2 and KCNQ3 genes. Weshow that both KCNQ2 and KCNQ3 contain GC box motifs andprovide evidence that their transcription is enhanced by the Sp1transcription factor. We also show that expression of bothKCNQ2 and KCNQ3 is repressed by REST (repressor element1-silencing transcription factor) and that expression of REST inneurons is sufficient to repress KCNQ2 and KCNQ3 expression,inhibiting functional expression of the M-current and resultingin hyperexcitable neurons. We show that REST levels are in-creased in dorsal root ganglia (DRG) neurons in response toinflammatory mediators and that Kv7.2 levels and M-currentdensity are reduced, suggesting a potential role in regulating in-flammatory pain responses. Neuronal expression of REST isincreased in response to sustained neuronal hyperactivity, i.e., in

Received April 19, 2010; revised June 15, 2010; accepted June 16, 2010.This work was supported by European Commission Grant LSHM-CT-2004-503038 (I.C.W.), Wellcome Trust Grants

080593/Z06/Z, 070740, and 080833/Z/06/Z, and Medical Research Council Grant 82507 (N.G.). We thank KatarzynaMarszalek for expert technical assistance and Dr. Mark Shapiro for the valuable comments on this manuscript.

The authors declare that they have no conflict of interest.*M.M. and L.O. contributed equally to this work.Correspondence should be addressed to Dr. Ian C. Wood, Institute of Membrane and Systems Biology, Faculty of

Biological Sciences, University of Leeds, Leeds, LS2 9JT, UK. E-mail: i.c.wood@leeds.ac.uk.M. Mucha’s present address: Department of Cell Physiology and Pharmacology, Faculty of Medicine and Biolog-

ical Sciences, University of Leicester, Leicester, LE1 9HN, UK.C. Dalle’s present address: Pain Therapeutics, Pfizer Global Research and Development, Sandwich, CT13 9NJ, UK.DOI:10.1523/JNEUROSCI.1981-10.2010

Copyright © 2010 the authors 0270-6474/10/3013235-11$15.00/0

The Journal of Neuroscience, October 6, 2010 • 30(40):13235–13245 • 13235

epileptic insults (Palm et al., 1998), cerebral ischemia (Calderoneet al., 2003), and a model of neuropathic pain (Uchida et al.); wetherefore suggest that, by repressing KCNQ2 and KCNQ3 expres-sion, REST may contribute to chronic overexcitability of neuro-nal circuits seen in epilepsy and chronic pain.

Materials and MethodsCell culture. SHSY-5Y and HEK293 cells were grown in DMEM/F-12supplemented with 10% fetal calf serum (PAA Laboratories), 6 g/literpenicillin, 10 g/liter streptomycin, and 2 mM L-glutamine at 37°C and 5%CO2. DRG neurons were cultured as described previously (Crozier et al.,2007; Linley et al., 2008; Liu et al., 2010); briefly, ganglia were extractedfrom 7- to 10-d-old rats from all spinal levels. Ganglia were enzymaticallydissociated in HBSS supplemented with collagenase type 1A (1.5 mg/ml)and dispase (15 mg/ml; Invitrogen) at 37°C for 15–20 min. Cells werethen mechanically triturated, washed twice by centrifugation (800 rpmfor 5 min), resuspended in 800 �l of growth medium, and plated ontoglass coverslips coated with poly-D-lysine and laminin. DRG neuronswere cultured for 2–5 d in a humidified incubator (37°C, 5% CO2) inDMEM supplemented with GlutaMAX I (Invitrogen), 10% fetal bovineserum, penicillin (50 U/ml), and streptomycin (50 �g/ml). For the in-flammatory conditions, cells were incubated with the following inflam-matory mediators for 48 h: 1 �M bradykinin, 1 �M histamine, 1 �M ATP,1 �M substance P, and 10 �M PAR-2 activating peptide. For patch-clampexperiments, cells were washed twice and media were replaced with freshmedia that did not contain inflammatory mediators for at least 2 h beforeexperiments, to remove any acute effects of inflammatory mediators onthe M-current.

DNase I hypersensitivity assay. SHSY-5Y cells were washed twice withice-cold PBS, resuspended in DNase I buffer (10 mM Tris, pH 8.0, 50 mM

KCl, 5 mM MgCl2, 3 mM CaCl2, 1 mM DTT, 0.1% Nonidet P-40, and 8%glycerol) and incubated on ice for 10 min. Cells were counted and trans-ferred to a Dounce homogenizer and lysed with 15 strokes of a type Bpestle. Nuclei were divided into independent samples (�1.6 � 10 5 perreaction) and treated with 0, 1, 3, 5, and 10 U of DNase I (Sigma) at 37°Cfor 5 min. Reactions were stopped by addition of equal volume of stop/lysis buffer (20 mM EDTA, 1% SDS, and 0.1 mg/ml proteinase K) andincubated at 55°C overnight. DNA was extracted several times with phe-nol/chloroform and once with chloroform and then ethanol precipi-tated. DNA was resuspended in water, and the concentration wasmeasured by spectrophotometry. Ten micrograms of each sample weredigested overnight with the appropriate restriction enzymes in a finalvolume of 30 �l and electrophoresed through 40 mM Tris-acetate, 1 mM

EDTA, pH 8, and 0.6% (w/v) agarose gel at 0.5 V/cm. DNA was trans-ferred to Hybond XL membranes (GE Healthcare) using alkaline capil-lary transfer. DNA fragments were radiolabeled using Prime-It kit(Stratagene) and [�- 32P]dATP. Hybridization to Hybond XL mem-branes was performed as per the instructions of the manufacturer.

Gel mobility shift assay. Nuclear extracts were prepared from SHSY-5Yand JTC-19 cells using the method described previously (Andrews andFaller, 1991). DNA probes were radioactively labeled by Klenow (NewEngland Biolabs) fill in of a restriction-enzyme-generated overhang using[�-32P]dATP (GE Healthcare) and incubated with protein extracts beforeelectrophoresis through a 4% nondenaturing acrylamide gel. Specificity ofDNA–protein binding was assessed by including 10 pmol of double-stranded oligonucleotides, or 1 �l of anti-Sp1 (2 mg/ml, H-255; Santa CruzBiotechnology) or 1 �l of anti-REST (2 mg/ml, P-18; Santa Cruz Biotech-nology) antibodies. Complimentary oligonucleotides with the following se-quences were annealed and used for competition experiments: consensusSp1, 5�-GATCCCGATCGGGGCGGGGCG-3�; nonspecific, 5�-AATTC-CCGCGAGGGGCGCCTAGTCCCCATG-3�; CHRM4 repressor element 1(RE1), 5�-GTACGGAGCTGTCCGAGGTGCTGAATCTGCCT-3�; humanKCNQ2 RE1, 5�-GATCCTGGTCAGGACCATGGCCAGCACCCC; dogKCNQ2 RE1, 5�-GATCTGCTGCTCAGGACCACGGCCAGCGCCTC;mouse KCNQ2 RE1, 5�-GATCTTGAGTCCAGGACCATGGTCAG-CACCAC; rat KCNQ2 RE1, 5�-GATCTTTGCGTCCAGGACCATGGT-CAGCGCCAC; KCNQ1 RE1, 5�-GATCCGGGCCTGCACCCAGG-ACAGGGCC; KCNQ3 RE1, 5�-GATCAGGCTCAGGACCTAGGACAGT-

TCC; KCNQ4, 5�-GATCCTGTCCAGGACCTGAGCCAGGGCT; andKCNQ5 RE1, 5�-GATCCTTGTCAGCACCTAGGACAGAGAT.

Luciferase assay. Fragments of human KCNQ2 gene corresponding to�64/�179 and �416/�151 bp regions were amplified by PCR with thefollowing primers: Sp1 sense (s), 5�-AAAAGATCTCATGGTGCCTG-GCGGGAGG; Sp1 antisense (a), 5�-AAGAGCTCAACGCGGGGC-GGGGCGAG; AscIs, 5�-GGCGCGCCTCGGGCTCAGGCTCAG; andAscIas, 5�-GGCGCGCCCTTGGTCCCTTCTGCC. Underlined nucleotidescorrespond to newly formed restriction sites. Additionally, we introducedmutations into the Sp1 binding site to prevent Sp1 binding using the primerfor Sp1sMut, 5�- AAGAGCTCAACGCGGGGCAAGGCGAG, containing asubstitution of GG with AA (in bold). Amplified fragments were digestedwith appropriate restriction enzymes and cloned into pGL3 Basic (Pro-mega). Cells were transfected in 24-well plates with 250 ng of pGL3 plasmidand 1 ng of pRL cytomegalovirus (Renilla luciferase; Promega) using 4 �l ofLipofectamine (Invitrogen). Each transfection was performed in triplicatefor each experiment. Cells were harvested 48 h after transfection into passivelysis buffer, and luciferase and Renilla luciferase expression were quantifiedusing a Dual Luciferase Assay kit (Promega) on a Mediators PHLluminometer.

Quantitative reverse transcriptase-PCR. Total RNA was extracted fromdorsal root ganglia cells using Tri-reagent (Sigma). DNase-treated (Ambion)RNA was reverse transcribed by using Moloney murine leukemia virus re-verse transcriptase (Promega) and purified using Qiaquick columns(Qiagen). Quantitative PCR was performed using SYBR Green incorpora-tion (Bio-Rad) for duplicate samples in each experiment. The specificity ofPCR was verified by melt curve analysis of products obtained from cDNA aswell as controls in which the reverse transcriptase was omitted. Levels ofsignal were normalized to levels of U6 small nuclear RNA, which were notdifferent between any of the datasets. Significance was tested using Student’st test. Primers used were as follows: U6s, 5�-CTCGCTTCGGCAGCACA;U6a, 5�-AACGCTTCACGAATTTGCGT; KCNQ2s, 5�-GTGGTCTA-CGCTCACAGCAA; KCNQ2a, 5�-AGGGTAAACGTCGCTGCTAA;KCNQ3s, 5�-GCCCACAGTCCTGCCCATCTTGAC; KCNQ3a, 5�-CCGT-TCCAGTTCCTCGTGGTTGACG; RESTs, 5�-CGAACTCACACAG-GAGAACG; RESTa, 5�-GAGGCCACATAATTGCACTG; SCN2As,5�-GCTGCATGTCTCTCTTGCTG; and SCN2Aa, 5�-GGACCGATTTGCT-TCACTTC.

Western blot and immunocytochemistry. For Western blots, 10 �g ofprotein extracts from control and mithramycin A-treated SHSY-5Y cellswere electrophoresed through an SDS/10% (w/v) polyacrylamide gel us-ing standard procedures, transferred to Hybond C-Extra (GE Health-care), and screened with 1:1000 dilution of anti-Kv7.2 (Alomone Labs)antibody and a 1:1000 dilution of horse radish peroxidase-conjugatedanti-rabbit antibody. Antibody staining was visualized using an ECL plusdetection kit (GE Healthcare). For immunocytochemistry, cells werefixed and permeabilized in acetone/methanol 1:1 for 5 min on ice, fol-lowed by 0.1 M PBS/0.1% Triton X-100 for 5 min at room temperatureand blocked with 10% donkey serum in PBS for 2 h at room temperature.Primary antibody incubation was performed in 10% donkey serum/0.4 M

PBS overnight at 4°C at 1:100 dilution for rabbit anti-REST (H290; SantaCruz Biotechnology), 1:1000 rabbit anti-Kv7.2 (kind gift from Dr. MarkShapiro, University of Texas Health Science Center at San Antonio, SanAntonio, TX) or 1:1000 guinea pig anti-TRPV1 (Neuromics), followedby 1:1000 Alexa Fluor-555 donkey anti-rabbit or Alexa Fluor-488 donkeyanti-guinea pig secondary antibody (Invitrogen) in 0.1 M PBS for 2 h atroom temperature. Coverslips were mounted in Vectashield with 4�,6�-diamidino-2-phenylindole (DAPI) and imaged using a Carl Zeiss LSM510inverted confocal microscope. Among comparisons between cell conditions,coverslips were mounted on the same slide and imaged with the same lasersettings. Image analysis was performed using NIH ImageJ software.

Plasmid and adenoviral delivery. The coding sequence for Sp1 was ampli-fied using the primers 5�-GCGAATTCAATGAGCGACCAAGATCACTCAand 5�-CGGAATTCTCAGAAACCATTGCCACTGAT and cloned into theexpression plasmid pTARGET (Promega). Freshly isolated DRG neuronswere transfected with pTARGET Sp1 or control plasmid using a Nucleofec-tor device (Amaxa) and used 3–5 d after transfection. The green fluorescentprotein (GFP) plasmid pmaxGFP (Amaxa) was cotransfected as a marker ofthe efficiency of transfection. The adenoviral constructs have been described

13236 • J. Neurosci., October 6, 2010 • 30(40):13235–13245 Mucha et al. • KCNQ Gene Regulation

previously (Wood et al., 2003). Green fluorescent protein was a marker of theefficiency of viral infection. Cells were cultured and infected with adenoviralparticles for 48 h before harvesting and RNA or electrophysiological analysis.

Electrophysiology. In patch-clamp experiments, the standard bath so-lution contained the following (in mM): 160 NaCl, 2.5 KCl, 2 CaCl2, 1MgCl2, and 10 HEPES, pH adjusted to 7.4 with NaOH. For perforated-patch experiments, the patch pipette contained 90 mM K-acetate, 20 mM

KCl, 1 mM CaCl2, 3 mM MgCl2, 3 mM EGTA, 40 mM HEPES, and 400�g/ml amphotericin B, pH adjusted to 7.4 with NaOH. Currents wereamplified using an EPC-10 patch-clamp amplifier (HEKA) and recordedusing Patchmaster software (version 2.2; HEKA). The current signal wassampled at 1 kHz and filtered online at 500 Hz using a software-basedBessel filter. Patch pipettes were fabricated from borosilicate glass(Harvard Apparatus) using a horizontal puller (DMZ-universal puller;Zeitz-Instrumente GmbH) and heat polished to a resistance of 2– 4 M�.Cells were mounted on an inverted microscope (TE-2000; Nikon) in alow-profile perfusion chamber fed by a gravity perfusion system flowingat �2 ml/min, resulting in a bath exchange time of �15 s. Series resis-tance was corrected online by up to 70% using the Patchmaster software,and liquid junction potentials were corrected. The magnitude of theneuronal M-current was measured from the deactivation current whenstepping the membrane voltage from �30 to �60 mV and was calculatedas the difference between current amplitude at 10 ms into the voltagepulse and that at the end of the pulse. This analysis method was designedto minimize any contribution from other K� currents and capacitanceartifacts. XE991 [10,10-bis(4-pyridinylmethyl)-9(10H)-anthracenone di-hydrochloride] was obtained from Tocris Bioscience. All analysis of patch-clamp data was conducted using Fitmaster software (version 2.11; HEKA).

Statistics. All data are given as mean � SEM. Differences betweengroups were assessed by Student’s t test or one-way ANOVA. The differ-ences were considered significant at p � 0.05.

ResultsIdentification of DNase I hypersensitive sites in the humanKCNQ2 geneTo identify regions of the KCNQ2 gene that are important forregulation of expression in vivo, we used a DNase I hypersensi-tivity assay. DNase I hypersensitive sites (HSS) correlate well withimportant transcriptional regulatory regions, such as promoters,enhancer, and repressor elements (Fritton et al., 1987; Steiner et al.,1987; Elgin, 1988; Wong et al., 1997), and map to regions that recruittranscription factors (Wong et al., 1995; Boyes and Felsenfeld,1996). A DNase I HSS assay was performed in SHSY-5Y cells,which are known to express KCNQ2 and KCNQ3 (Wickenden etal., 2008) and identified two HSS within the KCNQ2 gene (Fig.1A). We performed a bioinformatic analysis of the regions iden-tified to look for transcription factor binding sites. An analysis ofHSS1 identified a putative Sp1 transcription factor binding site(Fig. 1B) that is conserved within the promoters of the human,rat, and mouse KCNQ2 genes. It has been shown previously thatregions of DNA that recruit Sp1 display hypersensitivity to DNaseI (Philipsen et al., 1990; Pruzina et al., 1991; Jiang et al., 1997;Sinha and Fuchs, 2001). An analysis of the region surroundingHSS2 identified a putative binding site for the transcriptionalrepressor REST that is conserved within multiple mammalianspecies (Fig. 1C). REST binding to DNA is known to be sufficientto produce a DNase I HSS (Wood et al., 2003).

KCNQ2 and KCNQ3 genes contain similarregulatory elementsTo determine whether the potential Sp1 factor binding sites areimportant for KCNQ2 gene regulation, we used an in vitro DNA–protein interaction experiment to examine the recruitment ofSp1. A radiolabeled fragment of the KCNQ2 promoter contain-ing the Sp1 site was incubated with nuclear protein extracts fromSHSY-5Y cells. Incubation of this region of DNA with nuclear

protein produced two slowly migrating bands (Fig. 1D, lane 1,Sp1:DNA, Non spec) that could be competed by excess of anunlabeled Sp1 consensus sequence (Fig. 1D, lane 2). The fastermigrating DNA–protein complex could also be competed by anunrelated sequence (Fig. 1D, lane 3), indicating that this complexis the result of nonspecific protein–DNA interactions. The slowermigrating protein–DNA complex could only be competed with aconsensus Sp1 sequence and was further retarded by an anti-Sp1antibody (Fig. 1D, lane 4, Sp1:Ab:DNA), indicating the presenceof Sp1 protein within this complex. Binding of the radiolabeledprobe was also competed by an excess of unlabeled wild-type Sp1sequence from the KCNQ2 promoter (Sp1 WT) (Fig. 1D, lanes 5,6) but not by a KCNQ2 sequence containing mutations predictedto destroy the Sp1 site (Sp1 mut) (Fig. 1D, lanes 7, 8). These dataindicate that the Sp1 site within the KCNQ2 promoter can recruitSp1 protein present in SHSY-5Y cells. Given the overlap ofKCNQ2 and KCNQ3 expression and their shared evolutionaryorigin by duplication from a common ancestor gene (Hill et al.,2008), it is quite likely that these two genes would share at leastsome regulatory mechanisms. We therefore searched the KCNQ3gene for potential Sp1 and RE1 sites. We identified a conservedSp1 site in the KCNQ3 gene just upstream of the first exon, at anequivalent location to the Sp1 site in the KCNQ2 gene (Fig.1E,F). We also identified an RE1 site in the KCNQ3 gene that isconserved across multiple mammalian species (Fig. 1E,G). Thus,the two regulatory elements that we identified within the KCNQ2gene are also present in the KCNQ3 gene.

Sp1 regulates KCNQ2 and KCNQ3 expressionTo determine whether the Sp1 site is functionally important inregulating KCNQ2 and KCNQ3 transcription, we used a reportergene assay with the KCNQ promoter regions. We cloned a regionof the KCNQ2 promoter encompassing nucleotides �416/�151bp relative to the transcription start site into a luciferase reportervector such that luciferase expression is driven by KCNQ2 pro-moter sequences. This 568 bp region encompasses the Sp1 siteand 137 bp of the first exon. We also produced a truncated regionthat lacks some of the sequence upstream of the Sp1 site (�64/�179), and we introduced mutations into the Sp1 site that pre-vent Sp1 binding (�64/�179 Sp1 mut) (Fig. 1D, lanes 7, 8). Wetransfected these DNA constructs into SHSY-5Y cells and mea-sured the resulting luciferase activity. Luciferase activity from the�416/�151 bp promoter fragment was sixfold higher than thecontrol plasmid (Fig. 2A, compare �416/�151 bp with pGL3Basic), suggesting that this region of the KCNQ2 gene and the Sp1site within it elevate expression of KCNQ2. Removal of the up-stream region did not result in a significant loss of expression,suggesting that this region is not important in regulating tran-scription in SHSY-5Y cells (Fig. 2A, compare �416/�151 with�64/�179 bp). Introduction of mutations that prevent Sp1binding (see Materials and Methods) (Fig. 1D), however, re-sulted in loss of �50% luciferase activity (Fig. 2A, compare �64/�179 with �64/�179 bp Sp1 mut), consistent with the hypothesisthat KCNQ2 expression is regulated by Sp1. In our bioinformaticanalysis, we noticed that the KCNQ2 gene contained a second Sp1site that overlapped with the first although its score was less,suggestive of a lower affinity for Sp1 (supplemental Fig S1A,available at www.jneurosci.org as supplemental material). In gelmobility shift experiments, this second site did indeed bind toSp1, showing a lower affinity than our originally identified site(supplemental Fig S1B, available at www.jneurosci.org as supple-mental material). This second Sp1 site is unlikely to be bound bySp1 in the wild-type promoter because the presence of Sp1 at the

Mucha et al. • KCNQ Gene Regulation J. Neurosci., October 6, 2010 • 30(40):13235–13245 • 13237

high-affinity site, which overlaps with thelow-affinity site, would presumably oc-clude it. Mutation of the high-affinity site,however, would uncover this second site,and the low level of promoter activity seenfor the construct �64/�179 Sp1 mut islikely to be attributable to low levels of Sp1recruitment. To determine whether theKCNQ3 gene could also be regulated bySp1, we conducted a bioinformatic searchof the human KCNQ3 promoter region.We identified a sequence matching a con-sensus Sp1 site that was highly conservedbetween the human, mouse, and rat ge-nomes (Fig. 1F). To determine whetherthis potential Sp1 site is functional, wecloned an 876 bp region of the humanKCNQ3 gene encompassing nucleotides�540 to �336 relative to the transcriptionstart site and including the Sp1 site intothe same luciferase reporter vector as forthe KCNQ2 promoter. We also cloned aregion of the promoter encompassing nu-cleotides �169 to �336 that lacks theconserved Sp1 site. Luciferase activityfrom the �540/�336 promoter regiondrove high levels of luciferase expressionin SHSY-5Y cells (Fig. 2B), which was re-duced by �80% when the region contain-ing the Sp1 sites was removed (Fig. 2B).Together, these data suggest that Sp1 canregulate both KCNQ2 and KCNQ3 pro-moter activity.

Sp1 regulates the endogenous KCNQ2and KCNQ3 genesHaving shown that Sp1 can be recruited tothe KCNQ2 promoter in vitro and the Sp1site within the promoter is important fordriving reporter gene activity, we wantedto test the relevance of Sp1 transcriptionfactor activity to endogenous KCNQ2 andKCNQ3 expression. To do this, we tookadvantage of mithramycin A, a com-pound that binds to Sp1 binding sites (GCbox) and inhibits the function of Sp1 (Rayet al., 1989; Liu et al., 2002; Kim et al.,2006). Transfection of KCNQ2 andKCNQ3 luciferase reporter constructsinto SHSY-5Y cells in the presence ofmithramycin A resulted in a 66 and 90%reduction in luciferase activity, respec-tively (Fig. 2C). To determine the role ofSp1 on endogenous gene expression, we incubated neurons iso-lated from DRG in the presence of 200 nM of mithramycin A for48 h and measured mRNA expression using quantitative PCR. Inthe presence of mithramycin A, KCNQ2 and KCNQ3 levels werereduced by 90% (Fig. 2D), suggesting that Sp1 is critical for driv-ing endogenous KCNQ2 and KCNQ3 expression. Expression ofthe SCN2A gene, which encodes the type II sodium channel, wasunaffected by mithramycin A, suggesting that the effect onKCNQ expression was not the result of a global change in geneexpression (Fig. 2D). Incubation of SHSY-5Y cells with mithra-

mycin A also resulted in a reduction of Kv7.2 protein (Fig. 2E). Asan alternative approach to determine whether Sp1 was importantin regulating expression of KCNQ2 and KCNQ3, we transfectedDRG neurons with an Sp1 expression plasmid and measuredKv7.2 expression and M-current densities. We transfected a plas-mid expressing GFP alongside either the Sp1-expressing or thecontrol plasmid to identify transfected neurons (Fig. 3A). Inwhole-cell voltage-clamp experiments, IM was measured from astandard square voltage pulse protocol stepping to �60 mV froma holding potential of �30 mV (Fig. 3B, inset) as an amplitude of

Figure 1. DNase I hypersensitive sites present in the KCNQ2 gene. A, Top shows a schematic representation of the 5� of theKCNQ2 gene. Filled boxes represent exons, and the transcription start site is marked by an arrow. The location of identifiedregulatory elements (Sp1 and RE1) are shown as open boxes. Below shows two DNase I hypersensitivity assays. Nuclei wereisolated from SHSY-5Y cells and analyzed for DNase I hypersensitive sites. DNA was digested with HindIII (top) or AseI (bottom), andthe Southern blot was hybridized with a KCNQ2-specific probe. The positions of the hypersensitive sites (HSS1 and HSS2) areindicated on the diagram as is the position of the full-length DNA fragment (F). B, DNA sequence alignment of the Sp1 sequenceidentified in the HSS1 region of the human KCNQ2 gene with the equivalent regions of the mouse and rat KCNQ2 genes. Theconsensus Sp1 sequence is shown above the alignment (r, A or G; s, C or G). C, DNA sequence alignment of the RE1 sequenceidentified in the HSS2 region of the human KCNQ2 gene with the equivalent regions of the chimp, rhesus, mouse, rat, dog, cow andelephant KCNQ2 genes. The consensus RE1 sequence is shown above the alignment (n, A, C, G, or T; y, C or T; m, A or C). D, Gelmobility shift assay of the Sp1 region of the KCNQ2 promoter. Labeled DNA was incubated with protein isolated from SHSY-5Y cellsin the presence of no competitor (No comp, lane 1), oligonucleotides containing a consensus Sp1 sequence (Sp1 con, 1 �M, lane 2),an unrelated sequence (Non spec, 1 �M, lane 3), an anti-Sp1 antibody (�-Sp1, 1 �g, lane 4), or oligonucleotides containing thewild-type or mutated Sp1 sequence from the KCNQ2 promoter (Sp1 wt: 1 �M, lane 5; 10 �M, lane 6; Sp1 mut: 1 �M, lane 7; 10 �M,lane 8). Position of complexes containing Sp1 (Sp1:DNA), �-Sp1 (Ab:Sp1:DNA), and resulting from nonspecific binding (Non spec)are shown on the left. E, A schematic representation of the 5�of the KCNQ3 gene. Filled boxes represent exons, and the transcriptionstart site is marked by an arrow. The location of identified regulatory elements (Sp1 and RE1) are shown as open boxes. F, DNAsequence alignment of the Sp1 sequence identified in the human KCNQ3 gene with the equivalent regions of the mouse and ratKCNQ3 genes. The consensus Sp1 sequence is shown above the alignment (r, A or G; s, C or G). G, DNA sequence alignment of the RE1sequence identified in the human KCNQ3 gene with the equivalent regions of the chimp, rhesus, mouse, rat, dog, cow, andelephant KCNQ2 genes. The consensus RE1 sequence is shown above the alignment (n, A, C, G, or T; y, C or T; m, A or C).

13238 • J. Neurosci., October 6, 2010 • 30(40):13235–13245 Mucha et al. • KCNQ Gene Regulation

deactivating tail current at �60 mV sensitive to specific M channelblocker XE991 (3 �M) applied at the end of each recording.M-current density was significantly larger in Sp1-transfected com-pared with control neurons (Fig. 3B,C) (1.54 � 0.34 pA/pF, n 8compared with 0.91�0.12 pA/pF, p0.05). Consistently, the Kv7.2immunoreactivity was greater in Sp1-transfected than in control neu-rons(Fig.3A,D)(relativepixel intensityof9.53�2.15,n4comparedwith 5.49 � 0.85, n 3, p 0.05). Together, these data implicate animportant role for Sp1 in promoting both KCNQ2 and KCNQ3 geneexpression and enhancing levels of the M-current in neurons.

KCNQ genes are regulated by RESTBioinformatic analysis of the second DNase I hypersensitive site(Fig. 1A, HSS2) in the KCNQ2 gene identified a putative RE1binding site for REST. The sequence for this site is conservedbetween multiple mammalian species (Fig. 1C), suggesting that itis functionally important. REST was first identified as a repressorof neuronal gene expression (Chong et al., 1995; Schoenherr andAnderson, 1995) but also has roles in heart, blood vessels, andepithelial cells (Ooi and Wood, 2007). We used a gel mobilityshift assay to determine whether the RE1 site identified in theKCNQ2 gene can recruit REST protein. Incubation of a radioac-tive, RE1-containing, SCN2A promoter region with nuclear ex-tracts resulted in DNA–protein complexes (Fig. 4A, lane 1). Oneof these (indicated with an arrow on Fig. 4A) was competed by anRE1 sequence from the CHRM4 gene but not by an unrelated

sequence (Fig. 4A, lanes 2, 3). Further-more, an anti-REST antibody producedan additional shift of this complex, indi-cating the presence of REST protein (Fig.4A, lane 4). Oligonucleotides containingRE1 sequences from human, dog, mouse,and rat were all able to compete for RESTbinding, indicating that the KCNQ2 genecontains an evolutionary conserved, func-tional RE1 site (Fig. 4A, lanes 5–12).

To determine whether RE1 sites arepresent in other KCNQ genes, we used ourposition-specific scoring matrix (Johnsonet al., 2006) to scan each of the five KCNQgenes for the best match to an RE1. High-scoring RE1s were identified in KCNQ2,KCNQ3, and KCNQ5, whereas the closestsequences in KCNQ1 and KCNQ4 to anRE1 site had low scores with our matrix,suggesting they would not bind REST. Wetested each of these potential RE1 se-quences in a gel mobility shift assay to de-termine which would bind REST withhigh affinity (Fig. 4B). The RE1 sequencesfrom KCNQ2, KCNQ3, and KCNQ5 wereeach able to compete for REST binding,whereas the sequences in KCNQ1 andKCNQ4 could not (Fig. 4B). To deter-mine whether REST could regulate theKCNQ2 and KCNQ3 promoters, wecloned the RE1 sequences upstream of theluciferase reporter constructs (Fig. 2A)and transfected them into two cell lines,HEK293, which express robust levels offull-length REST natively, and SHSY-5Y,which express low levels of REST as well asa truncated version of REST resulting

from an alternative splice variant (our unpublished observa-tions). Inclusion of the RE1 site resulted in robust repression ofluciferase activity for both KCNQ2 and KCNQ3 promoters inHEK293 (Fig. 4C), indicating that recruitment of REST to theRE1 site represses KCNQ2 and KCNQ3 expression. Consistentwith the observation that SHSY-5Y cells express less REST pro-tein, inclusion of the RE1 site resulted in only modest repressionof luciferase activity for the KCNQ3 promoter in SHSY-5Y,whereas the KCNQ2 promoter did not show a significant reduc-tion in activity (Fig. 4D). REST is normally expressed in highlevels in non-neuronal cells and only at low levels in neurons (Sunet al., 2005; Olguín et al., 2006). Our luciferase data suggest thatthe absence or the low levels of REST in neurons may be impor-tant for permitting KCNQ2 and KCNQ3 expression. However,after epileptic seizures, global ischemia, or neuropathic injury,conditions characterized by periods of sustained neuronal hyper-activity, REST expression was seen in neurons (Palm et al., 1998;Calderone et al., 2003; Uchida et al., 2010). To determine whethersuch increased REST expression in neurons may affect KCNQ2and KCNQ3 expression, we infected cultured DRG neurons witheither GFP alone (control) or REST and GFP-expressing adeno-virus (Ad) particles. Infection of DRG neurons with control ad-enovirus had no effect on KCNQ2 mRNA, whereas infection withadenovirus driving REST expression resulted in a significant re-duction of KCNQ2 and KCNQ3 mRNA (Fig. 4E).

Figure 2. The KCNQ2 and KCNQ3 promoters contain functional Sp1 sequences. A, Regions of the KCNQ2 promoter were clonedupstream of luciferase and transfected into SHSY-5Y cells. Shown are normalized luciferase values expressed relative to emptyvector, pGL3 Basic (mean � SEM; n 3; *p 0.05). Base pair numbers are expressed relative to transcription start sites. B,Regions of the KCNQ3 promoter were cloned upstream of luciferase and transfected into SHSY-5Y cells. Shown are normalizedluciferase values expressed relative to empty vector (mean � SEM; n 3; *p 0.05). C, Luciferase reporter vector containing theKCNQ2 or KCNQ3 promoter regions or empty vector, pGL3 Basic (Basic), were transfected into SHSY-5Y cells that were subsequentlytreated with either 100 nM mithramycin A or water control. D, Reverse transcriptase-PCR analysis of KCNQ2, KCNQ3, and SCN2AmRNA levels in control and mithramycin A-treated cultures of rat dorsal root ganglia cells. Levels are normalized to U6 gene(mean � SEM; n 3; *p 0.05). E, Antibodies to Kv7.2 and �-actin were used to analyze respective protein levels in extractsfrom control or mithramycin A (Mith, 100 nM)-treated SHSY-5Y cells.

Mucha et al. • KCNQ Gene Regulation J. Neurosci., October 6, 2010 • 30(40):13235–13245 • 13239

Overexpression of REST reducesM-current density and changes firingproperties of DRG neuronsKCNQ2 and KCNQ3 encode the two sub-units Kv7.2 and Kv7.3, which togetherform a heterotetrameric potassium chan-nel that is believed to be the most abun-dant M-channel isoform in the CNS andperipheral nervous system (Delmas andBrown, 2005). IM is a slowly activating anddeactivating potassium current that pro-vides a brake for repetitive action poten-tial (AP) firing (Wang et al., 1998). Todetermine whether increased neuronalREST expression could downregulatethe endogenous M-current, we infectedcultured DRG neurons with REST-expressing adenovirus and examined theelectrophysiological properties of theseneurons. DRG neurons significantly differin their function and channel expressionprofile; thus, to restrict our study to amore homogeneous population of neu-rons, we applied the TRPV1 agonist cap-saicin (1 �M) at the end of each recordingand only considered those neurons thatwere TRPV1 positive (nociceptive neu-rons). Whole-cell currents were measuredfrom small (whole-cell capacitance of23.7 � 2.8 pF, n 28) DRG neurons us-ing a standard voltage protocol forM-current (Fig. 5A, inset). In control cellsinfected with adenoviral construct codingfor GFP alone, the hyperpolarizing test pulse resulted in a slowlydeactivating whole-cell current, characteristic of M-current,which was sensitive to the specific M-channel blocker XE991 (3�M; 62 � 4% inhibition of deactivation current, n 8). In neu-rons infected with AdREST, IM was dramatically reduced to 14%of the value of control Ad neurons ( p 0.01, n 9 for GFP andn 11 for GFP � REST-expressing neurons) (Fig. 5C). We alsocompared responses of virally infected neurons with 1 �M of theTRPV1 agonist capsaicin (measured as the inward current at �60mV). Capsaicin responses were unaltered by REST overexpres-sion (control, �21.8 � 8.9 pA/pF, n 9; REST, �26.6 � 7.5pA/pF, n 11; ANOVA, p NS) (data not shown), suggestingthat the effect of REST was specific to its target genes. To furtherconfirm that viral infection did not affect IM nonspecifically,whole-cell currents were measured from non-infected neuronsfrom AdREST-exposed neuronal cultures. In these cells, IM wasnot significantly different from Ad controls (2.21 � 0.26 vs2.37 � 0.48 pA/pF, n 8; p 0.05) but was significantly largerthan REST-infected cells (0.31 � 0.10 pA/pF, n 8; ANOVA, p 0.01) (Fig. 5C). Thus, increasing REST expression (Fig. 5A, Ad-REST) resulted in almost a complete absence of the M-currentwhen compared with uninfected neurons from the same dish[Fig. 5A, AdREST (non-infected)] or neurons infected with acontrol adenovirus (Fig. 5A, Ad).

M-current is known to contribute to the resting membranepotential and acts as a brake on AP firing of DRG neurons (Pass-more et al., 2003; Linley et al., 2008). We therefore tested whetherincreasing REST expression affected the AP firing properties andthe resting membrane potential of DRG neurons. Using whole-cell current clamp, the membrane potential was adjusted to �65

mV by injection of current. Injection of �400 pA produced oneAP in the majority (seven of nine) of Ad-infected cells and twoaction potentials in the remainder (two of nine) of cells with asimilar pattern of firing observed in non-infected AdREST cells(five of six cells firing one AP). In contrast in REST overexpress-ing cells, 6 of 11 cells fired multiple action potentials (mean APnumber of 6.5), which did not show accommodation (Fig. 5B),indicating that the brake on neuronal firing had been removed.REST-overexpressing neurons also were significantly depolar-ized compared with controls (Fig. 5D: AdREST, �60.5 � 2.4 mV;Ad, �67.6 � 2.6 mV; Non-infected, �68.6 � 2.3 mV; n 8;ANOVA, p 0.01), consistent with a reduction in M-current inthese cells. REST has many gene targets (Bruce et al., 2004; John-son et al., 2009) in addition to the KCNQ genes, and the change inthe excitability profile of a DRG neuron after REST overexpression islikely to reflect multiple changes in the expression of ion channels atthe neuronal membrane. However, experimental data (Gamper etal., 2006) and modeling (Zaika et al., 2006) suggest that inhibition ofM-current alone is sufficient to explain the increase in DRG neuronfiring rate observed here. In additional support to this assumption,the selective Kv7 blocker XE991 produced an increase in AP firingwhen acutely applied to a nontransfected DRG neuron (Fig. 5E), theincrease in excitability was similar to that produced by REST over-expression. These data suggest that the regulation of KCNQ geneexpression levels underlies the increased neuronal excitability result-ing from expression of REST.

Although its levels are normally low, neuronal REST expres-sion was shown to increase after extended periods of neuronaloveractivity seen in seizures, ischemia, and neuropathic pain(Palm et al., 1998; Calderone et al., 2003; Uchida et al., 2010).

Figure 3. Sp1 enhances Kv7.2 expression and M-current density. A, Fluorescent images of DRG neurons transfected with aGFP-expressing plasmid and an Sp1-expressing (top) or control (bottom) plasmid. Transfected cells were identified by GFP expres-sion (green). Kv7.2 levels were analyzed using an anti-Kv7.2 antibody and Alexa Fluor-555-conjugated secondary antibody (red).Nuclei are counterstained with DAPI (blue). Scale bars, 10 �m. B, Whole cell voltage-clamp traces obtained by 500 ms voltagepulses from �30 to �60 mV as indicated by the voltage pulse protocol (inset). Black trace represents steady state basal current;red trace represents current in the presence of the specific M-channel blocker XE991 (3�M). Dotted line indicates zero current. C,Bars show pooled current density data determined from the DRG neurons transfected with an Sp1-expressing (Sp1; n8) or control (Con;n7) plasmid. Magnitude of M-current was calculated as the XE991-sensitive deactivation current when stepping from�30 to�60 mVand normalized to cell capacitance (IM density, mean � SEM; *p 0.05). D, Bars show pixel intensity of Kv7.2 staining in DRG neuronstransfected with Sp1-expressing (Sp1; n 4) or control (Con; n 3) plasmid (mean � SEM; *p 0.05).

13240 • J. Neurosci., October 6, 2010 • 30(40):13235–13245 Mucha et al. • KCNQ Gene Regulation

Having shown that REST can regulate the expression of KCNQ2and KCNQ3 in DRG neurons, we sought to determine whetherREST levels in DRG neurons may be increased physiologically.We recently reported that inflammatory mediators increase theexcitability of nociceptive DRG neurons via inhibition of theM-current mediated through short-term G-protein signalingevents (Linley et al., 2008; Liu et al., 2010). Long-term changes innociceptive neurons have been proposed to be important in thedevelopment of chronic pain (for review, see Basbaum et al.,2009; Linley et al., 2010) and are likely to be the result of changesin gene expression. We therefore cultured DRG neurons in thepresence of a mix of inflammatory mediators to mimic inflam-mation (1 �M bradykinin, 1 �M histamine, 1 �M ATP, 1 �M

substance P, and 10 �M PAR-2 activating peptide) and analyzedthe expression of REST and Kv7.2 specifically in the nociceptive,TRPV1-positive, neurons (Fig. 6A–C). In control conditions,REST levels in nociceptive neurons were very low, but they in-creased significantly in response to inflammatory mediators (Fig.6A,C). The levels of TRPV1 in nociceptive neurons did notchange significantly (Fig. 6C) (TRPV1 is not a predicted targetgene of REST). TRPV1-positive neurons cultured in inflamma-tory conditions showed lower Kv7.2 immunoreactivity thanthose in control conditions (Fig. 6B,C). We also analyzedM-current density in capsaicin-responsive neurons. For patchclamp, inflammatory mediators were washed out for at least 2 h

before experiments to remove any acuteeffects on IM. Inflammatory conditionsdramatically reduced IM compared withcontrol (0.25 � 0.04 compared with0.95 � 0.16 pA/pF; p � 0.01; n 9 forcontrol and n 5 for inflammatory), con-sistent with a downregulation of KCNQ2and KCNQ3. These data are consistentwith a mechanism whereby REST levelsare increased in nociceptive neurons inresponse to inflammatory signals, re-sulting in an increase in excitabilitybrought about by the reduced expres-sion of KCNQ2 and KCNQ3. Such amechanism could contribute to a long-term peripheral sensitization of in-flamed sensory fibers.

DiscussionIn this work, we have identified two com-mon mechanisms that regulate expressionof KCNQ2 and KCNQ3. We have identi-fied consensus sequences for the Sp1 tran-scription factor in each of the promoterregions. These sequences are evolutionar-ily conserved, and we show that Sp1 di-rectly interacts with the KCNQ2 promoterin vitro and that removal or mutation ofthese sequences results in reduced activityof the promoters. Furthermore, chemicalinhibition of Sp1–DNA interactions re-sult in reduced KCNQ2 and KCNQ3mRNA expression, whereas ectopic Sp1results in increased M-current and Kv7.2expression in cultured neurons. Our dataimplicating Sp1 as an important positive-acting transcription factor that regulatesKCNQ2 expression are consistent with aprevious report on the human KCNQ2

promoter (Xiao et al., 2001). Xiao et al. used a bioinformaticanalysis to identify GC boxes in the proximal promoter ofKCNQ2. Our new data show that these GC box regions are evo-lutionarily conserved and are also present in the KCNQ3 pro-moter. Furthermore, Sp1 does in fact interact with these GCboxes, and this interaction is important for expression of theendogenous KCNQ2 and KCNQ3 mRNA in neurons. The phys-iological relevance of Sp1 regulation of KCNQ2 and KCNQ3 re-mains to be determined, and, although Sp1 is often thought of asa constitutive transcriptional activator, recent evidence suggeststhat Sp1 activity is important for mediating changes in neuronalgene expression in response to developmental or disease stimuli.For example, induced expression of the reelin gene during neu-ronal differentiation is dependent on Sp1, and mutation in thebinding site for Sp1 within the reelin promoter is sufficient toprevent increased reelin expression (Chen et al., 2007). Sp1 is alsoimportant for the increased expression of the damage-inducedneuronal endopeptidase (DINE) gene in response to nerve injury(Kiryu-Seo et al., 2008). In this context, Sp1 acts as a platform towhich other transcription factors are recruited in response tonerve injury and activate expression of DINE (Kiryu-Seo et al.,2008).

Our data also highlight a second conserved mechanism for theregulation of KCNQ2 and KCNQ3: repression by REST. Both

Figure 4. The KCNQ2 and KCNQ3 promoters are regulated by the transcriptional repressor REST. A, B, Labeled DNA containingthe SCN2A RE1 site was incubated with nuclear protein in the presence of no competitor (No comp, lane 1), oligonucleotidescontaining an RE1 site from the CHRM4 promoter (CHRM4, lane 2), an unrelated sequence (Non spec, lane 3), an anti-REST antibody(�-REST, lane 4), and oligonucleotides containing sequence from the human (lanes 5, 6), dog (lanes 7, 8), mouse (lanes 9, 10), andrat (lanes 11, 12) KCNQ2 genes (A) or oligonucleotides containing potential RE1 sequences from the human KCNQ1–KCNQ5 genes(lanes 5–14) (B). Concentration of competing oligonucleotides were 1 �M (lanes 5, 7, 9, 11, 13) or 10 �M (lanes 2, 3, 6, 8, 10, 12,14). Positions of REST–DNA and REST–DNA–antibody complexes are identified by arrows on the left. C, D, Luciferase activity drivenfrom KCNQ2 and KCNQ3 promoters with or without the RE1 site, transfected into HEK293 cells (C) or SHSY-5Y cells (D). The arrowson the KCNQ3 RE1 site represent the direction of the sequence. Expression levels were normalized for transfection efficiency withRenilla luciferase and are expressed relative to empty plasmid, pGL3 basic (mean � SEM; n 3; *p 0.05). E, Reversetranscriptase-PCR analysis of KCNQ2 and KCNQ3 mRNA levels in cultured rat dorsal root ganglia cells infected with control adeno-virus or adenovirus expressing full-length REST (�). Data are normalized to U6 gene and expressed relative to cells infected withcontrol adenovirus (mean � SEM; n 3; *p 0.05).

Mucha et al. • KCNQ Gene Regulation J. Neurosci., October 6, 2010 • 30(40):13235–13245 • 13241

KCNQ2 and KCNQ3 contain evolution-arily conserved REST binding sites thatcan interact with REST in vitro, and over-expression of REST in neurons results inreduced KCNQ2 and KCNQ3 mRNA andloss of M-current. Interestingly, we alsoidentified a functional binding site forREST in the KCNQ5 gene, suggesting thatKCNQ5 may also be repressed by REST.Like KCNQ2 and KCNQ3, KCNQ5 iswidely expressed throughout the nervoussystem, and it has been suggested thatKv7.5 subunits may contribute to theM-current in vivo either on their own or incomplex with Kv7.3 (Lerche et al., 2000;Schroeder et al., 2000). In a whole genomechromatin immunoprecipitation analysisto look at REST binding in human Jurkatcells, REST is enriched at the RE1 sites forKCNQ2 and KCNQ3 (Johnson et al.,2007). In that study, REST was not foundto be present at the KCNQ5 RE1 site weidentified, but this may be a cell-specificeffect because enrichment of REST at theKCNQ5 RE1 site has been identified inseveral other human cell lines (Bruce etal., 2009). Although REST was originallyproposed to be important in silencingneuronal-specific genes in non-neuronalcells (Chong et al., 1995; Schoenherr andAnderson, 1995), it clearly has a func-tional role within the nervous system.REST is expressed at low levels in neuronsand regulates some target genes (Wood etal., 2003). Neuronal REST expression isupregulated in response to epileptic in-sults and cerebral ischemia resulting inreduced expression of brain derived neu-rotrophic factor (BDNF) and the gluta-mate receptor subunit gene GRIA2 andincreased expression of the substance Pgene PPT-A (Palm et al., 1998; Calderoneet al., 2003; Garriga-Canut et al., 2006;Spencer et al., 2006). In addition, nuclearlevels of REST protein are elevated in neu-rons in Huntington’s disease, leading toreduced BDNF expression (Zuccato et al.,2003, 2007).

Individuals who suffer an epileptic sei-zure become more susceptible to addi-tional seizure activity (Hauser and Lee,2002). The molecular mechanisms un-derlying this long-term change are notknown, although, like other features ofthe brain that show long-lasting changes,e.g., long-term potentiation, they mostlikely involve changes in gene expression.The expression of several transcriptionfactors has been shown to be increased inresponse to seizure activity in rodents.These include c-Fos (Morgan et al., 1987;Hiscock et al., 2001), nuclear factor-�B,Ap-1 (Rong and Baudry, 1996), DREAM

Figure 5. REST expression inhibits M-current and increases excitability in small-diameter DRG neurons. DRG neurons wereinfected with an adenoviral construct expressing GFP only (Ad) or REST and GFP (AdREST) and assessed functionally using whole-cell patch clamp. In cultures infected with REST and GFP, both green (AdREST) and non-green (AdREST non-infected) cells weretested. Only small-diameter neurons (18 –35 pF) responsive to the TRPV1 agonist capsaicin (1 �M) were investigated. A, Whole-cell voltage-clamp traces obtained by 500 ms voltage pulses from �30 to �60 mV as indicated by the voltage pulse protocol(inset). Black trace represents steady-state basal current; red trace represents current in the presence of the specific M-channelblocker XE991 (3 �M). Dotted line indicates zero current. B, Whole-cell current-clamp traces in which voltage was adjusted to �65mV by current injection and 4 s square current pulses to different test currents were applied (inset). C, Bars show pooled currentdensity data determined from the experiments shown in A. Magnitude of M-current was calculated as the XE991-sensitive deac-tivation current when stepping from �30 to �60 mV and normalized to cell capacitance (IM density). Number of experiments isindicated within bars [**p 0.01 compared to Ad and AdRes� (noninfected)]. D, Bars show mean resting membrane potentialsof GFP, REST, and non-infected neurons. Membrane potential was measured in current-clamp mode (0 pA) [**p 0.01 comparedto Ad and AdRest (noninfected)]. E, Exemplary current-clamp experiment showing excitatory effect of 3 �M XE991. Currentinjections at 400 pA (inset) were applied to whole-cell current-clamped DRG neurons before and during XE991 application. Novoltage adjustment was performed.

13242 • J. Neurosci., October 6, 2010 • 30(40):13235–13245 Mucha et al. • KCNQ Gene Regulation

(downstream regulatory element antagonist modulator) (Matsu-ura et al., 2002), and REST (Palm et al., 1998). One report has alsoshown that Sp1 levels increase after seizure (Feng et al., 1999),although another found no changes in levels of Sp1 (Rong andBaudry, 1996). Given our findings, an increase in expression ofthe transcriptional REST would result in a decrease in KCNQ2and KCNQ3 expression, resulting in an increase in the excitabilityof the affected neurons. REST represses gene expression by re-cruiting protein complexes that modify the posttranslationalmarks within chromatin (Ooi and Wood, 2007). Thus, anychanges induced on REST recruitment have the potential tobe stably maintained even if REST expression is subsequentlyreduced, providing a prospective mechanism for long-termchanges in gene expression levels. In addition to repressing tran-scription by modifying chromatin, REST has also been shown tointeract with Sp1 and inhibit the ability of Sp1 to enhance tran-

scription (Plaisance et al., 2005). Such amodel is consistent with our data in whichthe ability of REST to repress KCNQ2 andKCNQ3 appears to be dominant over theability of Sp1 to activate these genes andsuggests that interaction between RESTand Sp1 may be important for regulatingthe expression of many neuronal-specificgenes. Consistent with such a hypothe-sis is the observation that, in additionto KCNQ2 and KCNQ3, many REST-regulated genes are also known to be regu-lated by Sp1. For example, Chrm4 (Wood etal., 1996), Nmdar1 (Bai et al., 1998), Gria2(Myers et al., 1998), synatophysin (Lietz etal., 2003), and Bsx (Park et al., 2007) genesare all regulated by both REST and Sp1,whereas regulation of the �-opioid receptorgene involves an interaction with REST andthe Sp1 family member Sp3 (Kim et al.,2006).

Recently, much interest has been placedon Kv7 channels as a potential target for an-algesics in pain (Passmore et al., 2003; Linleyet al., 2008; Brown and Passmore, 2009;Wickenden and McNaughton-Smith,2009). Activators of Kv7 channels had ananalgesic effect in a hindpaw model ofchronic pain (Passmore et al., 2003), inhib-ited ectopic firing of axotomized sensory fi-bers (Roza and Lopez-Garcia, 2008), andreduced behavior associated with visceralpain (Hirano et al., 2007). Furthermore,pharmacological inhibition of M-channelsat peripheral nerve terminals by acute injec-tion of M-channel blocker XE991 producedspontaneous pain (Linley et al., 2008; Liu etal., 2010). Our data suggest that REST ex-pression increases whereas Kv7.2 expres-sion and M-current density decreases inDRG neurons under chronic inflamma-tory treatment. All these findings sug-gest that transcriptional downregulationof M-channel expression in nociceptivepathways may result in painful phenotypeand contribute to chronic pain condi-tions. Moreover, peripheral axotomiza-

tion in a rat model for neuropathic pain resulted in changes in thegene expression profile of DRGs (Xiao et al., 2002). Levels ofneither KCNQ nor REST genes were reported in this study, buttwo well known REST target genes, chromagranin A and SNAP25(synaptosomal-associated protein 25), showed reduced expres-sion in axotomized DRGs (Xiao et al., 2002). Recently, expressionof REST was shown to be upregulated after neuropathic injury(Uchida et al., 2010), thus it is tempting to speculate that theKCNQ–REST pathway discovered here may contribute to neuro-pathic pain pathophysiology.

Here we identify two transcription factors, Sp1 and REST, thatare important for regulating expression of the potassium channelgenes KCNQ2 and KCNQ3. Regulation of KCNQ2 and KCNQ3by these transcription factors may be important for long-termchanges in expression levels in response to epileptic seizure activ-ity and during chronic pain syndromes.

Figure 6. Inflammatory stimuli results in an increase in neuronal REST levels. A, B, Fluorescent images of DRG neurons culturedin control conditions (top) or exposed to a mixture of inflammatory mediators (bottom) for 48 h. REST (A, red), Kv7.2 (B, red), andTRPV1 (green) levels were analyzed by anti-REST, anti-Kv7.2, and anti-TRPV1 antibodies and Alexa Fluor-555 (REST and Kv7.2) orAlexa Fluor-488 (TRPV1) conjugated secondary antibodies. Scale bars, 10 �m. C, Bars show the pixel intensity of TRPV1, REST, andKv7.2 staining in TRPV1-positive neurons in control conditions (Con) and after exposure to inflammatory mediators (Inflam)(mean � SEM; n � 12; *p 0.05).

Mucha et al. • KCNQ Gene Regulation J. Neurosci., October 6, 2010 • 30(40):13235–13245 • 13243

ReferencesAndrews NC, Faller DV (1991) A rapid micropreparation technique for ex-

traction of DNA-binding proteins from limiting numbers of mammaliancells. Nucleic Acids Res 19:2499.

Bai G, Norton DD, Prenger MS, Kusiak JW (1998) Single-stranded DNA-binding proteins and neuron-restrictive silencer factor participate in cell-specific transcriptional control of the NMDAR1 gene. J Biol Chem273:1086 –1091.

Basbaum AI, Bautista DM, Scherrer G, Julius D (2009) Cellular and molec-ular mechanisms of pain. Cell 139:267–284.

Biervert C, Steinlein OK (1999) Structural and mutational analysis ofKCNQ2, the major gene locus for benign familial neonatal convulsions.Hum Genet 104:234 –240.

Boyes J, Felsenfeld G (1996) Tissue-specific factors additively increase theprobability of the all-or-none formation of a hypersensitive site. EMBO J15:2496 –2507.

Brown DA, Adams PR (1980) Muscarinic suppression of a novel voltage-sensitive K � current in a vertebrate neurone. Nature 283:673– 676.

Brown DA, Passmore GM (2009) Neural KCNQ (Kv7) channels. Br J Phar-macol 156:1185–1195.

Bruce AW, Donaldson IJ, Wood IC, Yerbury SA, Sadowski MI, Chapman M,Gottgens B, Buckley NJ (2004) Genome-wide analysis of repressor ele-ment 1 silencing transcription factor/neuron-restrictive silencing factor(REST/NRSF) target genes. Proc Natl Acad Sci U S A 101:10458 –10463.

Bruce AW, Lopez-Contreras AJ, Flicek P, Down TA, Dhami P, Dillon SC,Koch CM, Langford CF, Dunham I, Andrews RM, Vetrie D (2009)Functional diversity for REST (NRSF) is defined by in vivo binding affin-ity hierarchies at the DNA sequence level. Genome Res 19:994 –1005.

Calderone A, Jover T, Noh KM, Tanaka H, Yokota H, Lin Y, Grooms SY,Regis R, Bennett MV, Zukin RS (2003) Ischemic insults derepress thegene silencer REST in neurons destined to die. J Neurosci 23:2112–2121.

Charlier C, Singh NA, Ryan SG, Lewis TB, Reus BE, Leach RJ, Leppert M(1998) A pore mutation in a novel KQT-like potassium channel gene inan idiopathic epilepsy family. Nat Genet 18:53–55.

Chen Y, Kundakovic M, Agis-Balboa RC, Pinna G, Grayson DR (2007) In-duction of the reelin promoter by retinoic acid is mediated by Sp1. J Neu-rochem 103:650 – 665.

Chong JA, Tapia-Ramírez J, Kim S, Toledo-Aral JJ, Zheng Y, Boutros MC,Altshuller YM, Frohman MA, Kraner SD, Mandel G (1995) REST: amammalian silencer protein that restricts sodium channel gene expres-sion to neurons. Cell 80:949 –957.

Crozier RA, Ajit SK, Kaftan EJ, Pausch MH (2007) MrgD activation inhibitsKCNQ/M-currents and contributes to enhanced neuronal excitability.J Neurosci 27:4492– 4496.

Delmas P, Brown DA (2005) Pathways modulating neural KCNQ/M (Kv7)potassium channels. Nat Rev Neurosci 6:850 – 862.

Elgin SC (1988) The formation and function of DNase I hypersensitive sitesin the process of gene activation. J Biol Chem 263:19259 –19262.

Feng Z, Chang RC, Bing G, Hudson P, Tiao N, Jin L, Hong JS (1999) Long-term increase of Sp-1 transcription factors in the hippocampus afterkainic acid treatment. Brain Res Mol Brain Res 69:144 –148.

Fritton HP, Igo-Kemenes T, Nowock J, Strech-Jurk U, Theisen M, Sippel AE(1987) DNase I-hypersensitive sites in the chromatin structure of thelysozyme gene in steroid hormone target and non-target cells. Biol ChemHoppe Seyler 368:111–119.

Gamper N, Zaika O, Li Y, Martin P, Hernandez CC, Perez MR, Wang AY,Jaffe DB, Shapiro MS (2006) Oxidative modification of M-type K �

channels as a mechanism of cytoprotective neuronal silencing. EMBO J25:4996 –5004.

Garriga-Canut M, Schoenike B, Qazi R, Bergendahl K, Daley TJ, Pfender RM,Morrison JF, Ockuly J, Stafstrom C, Sutula T, Roopra A (2006) 2-Deoxy-D-glucose reduces epilepsy progression by NRSF-CtBP-dependent metabolicregulation of chromatin structure. Nat Neurosci 9:1382–1387.

Gribkoff VK (2008) The therapeutic potential of neuronal K V 7 (KCNQ)channel modulators: an update. Expert Opin Ther Targets 12:565–581.

Hauser WA, Lee JR (2002) Do seizures beget seizures? Prog Brain Res135:215–219.

Hill AS, Nishino A, Nakajo K, Zhang G, Fineman JR, Selzer ME, Okamura Y,Cooper EC (2008) Ion channel clustering at the axon initial segment andnode of Ranvier evolved sequentially in early chordates. PLoS Genet4:e1000317.

Hirano K, Kuratani K, Fujiyoshi M, Tashiro N, Hayashi E, Kinoshita M

(2007) Kv7.2–7.5 voltage-gated potassium channel (KCNQ2–5) opener,retigabine, reduces capsaicin-induced visceral pain in mice. Neurosci Lett413:159 –162.

Hiscock JJ, Mackenzie L, Medvedev A, Willoughby JO (2001) Kainic acidand seizure-induced Fos in subtypes of cerebrocortical neurons. J Neuro-sci Res 66:1094 –1100.

Jentsch TJ (2000) Neuronal KCNQ potassium channels: physiology androle in disease. Nat Rev Neurosci 1:21–30.

Jiang JG, Chen Q, Bell A, Zarnegar R (1997) Transcriptional regulation ofthe hepatocyte growth factor (HGF) gene by the Sp family of transcriptionfactors. Oncogene 14:3039 –3049.

Johnson DS, Mortazavi A, Myers RM, Wold B (2007) Genome-wide map-ping of in vivo protein-DNA interactions. Science 316:1497–1502.

Johnson R, Gamblin RJ, Ooi L, Bruce AW, Donaldson IJ, Westhead DR,Wood IC, Jackson RM, Buckley NJ (2006) Identification of the RESTregulon reveals extensive transposable element-mediated binding site du-plication. Nucleic Acids Res 34:3862–3877.

Johnson R, Samuel J, Ng CK, Jauch R, Stanton LW, Wood IC (2009) Evo-lution of the vertebrate gene regulatory network controlled by the tran-scriptional repressor REST. Mol Biol Evol 26:1491–1507.

Kim CS, Choi HS, Hwang CK, Song KY, Lee BK, Law PY, Wei LN, Loh HH(2006) Evidence of the neuron-restrictive silencer factor (NRSF) inter-action with Sp3 and its synergic repression to the mu opioid receptor(MOR) gene. Nucleic Acids Res 34:6392– 6403.

Kiryu-Seo S, Kato R, Ogawa T, Nakagomi S, Nagata K, Kiyama H (2008)Neuronal injury-inducible gene is synergistically regulated by ATF3,c-Jun, and STAT3 through the interaction with Sp1 in damaged neurons.J Biol Chem 283:6988 – 6996.

Lerche C, Scherer CR, Seebohm G, Derst C, Wei AD, Busch AE, Steinmeyer K(2000) Molecular cloning and functional expression of KCNQ5, a potas-sium channel subunit that may contribute to neuronal M-current diver-sity. J Biol Chem 275:22395–22400.

Lietz M, Hohl M, Thiel G (2003) RE-1 silencing transcription factor (REST)regulates human synaptophysin gene transcription through an intronicsequence-specific DNA-binding site. Eur J Biochem 270:2–9.

Linley JE, Rose K, Patil M, Robertson B, Akopian AN, Gamper N (2008)Inhibition of M current in sensory neurons by exogenous proteases: asignaling pathway mediating inflammatory nociception. J Neurosci28:11240 –11249.

Linley JE, Rose K, Ooi L, Gamper N (2010) Understanding inflammatorypain: ion channels contributing to acute and chronic nociception.Pflugers Arch 459:657– 669.

Liu B, Linley JE, Du X, Zhang X, Ooi L, Zhang H, Gamper N (2010) Theacute nociceptive signals induced by bradykinin in rat sensory neuronsare mediated by inhibition of M-type K � channels and activation ofCa 2�-activated Cl� channels. J Clin Invest 120:1240 –1252.

Liu W, Wang G, Yakovlev AG (2002) Identification and functional analysisof the rat caspase-3 gene promoter. J Biol Chem 277:8273– 8278.

Matsu-ura T, Konishi Y, Aoki T, Naranjo JR, Mikoshiba K, Tamura TA(2002) Seizure-mediated neuronal activation induces DREAM gene ex-pression in the mouse brain. Brain Res Mol Brain Res 109:198 –206.

Morgan JI, Cohen DR, Hempstead JL, Curran T (1987) Mapping patterns ofc-fos expression in the central nervous system after seizure. Science237:192–197.

Myers SJ, Peters J, Huang Y, Comer MB, Barthel F, Dingledine R (1998)Transcriptional regulation of the GluR2 gene: neural-specific expression,multiple promoters, and regulatory elements. J Neurosci 18:6723– 6739.

Olguín P, Oteíza P, Gamboa E, Gomez-Skarmeta JL, Kukuljan M (2006) RE-1silencer of transcription/neural restrictive silencer factor modulates ectoder-mal patterning during Xenopus development. J Neurosci 26:2820–2829.

Ooi L, Wood IC (2007) Chromatin crosstalk in development and disease:lessons from REST. Nat Rev Genet 8:544 –554.

Palm K, Belluardo N, Metsis M, Timmusk T (1998) Neuronal expression ofzinc finger transcription factor REST/NRSF/XBR gene. J Neurosci18:1280 –1296.

Park SY, Kim JB, Han YM (2007) REST is a key regulator in brain-specifichomeobox gene expression during neuronal differentiation. J Neuro-chem 103:2565–2574.

Passmore GM, Selyanko AA, Mistry M, Al-Qatari M, Marsh SJ, Matthews EA,Dickenson AH, Brown TA, Burbidge SA, Main M, Brown DA (2003)KCNQ/M currents in sensory neurons: significance for pain therapy.J Neurosci 23:7227–7236.

13244 • J. Neurosci., October 6, 2010 • 30(40):13235–13245 Mucha et al. • KCNQ Gene Regulation

Philipsen S, Talbot D, Fraser P, Grosveld F (1990) The beta-globin domi-nant control region: hypersensitive site 2. EMBO J 9:2159 –2167.

Plaisance V, Niederhauser G, Azzouz F, Lenain V, Haefliger JA, Waeber G,Abderrahmani A (2005) The repressor element silencing transcriptionfactor (REST)-mediated transcriptional repression requires the inhibi-tion of Sp1. J Biol Chem 280:401– 407.

Pruzina S, Hanscombe O, Whyatt D, Grosveld F, Philipsen S (1991) Hyper-sensitive site 4 of the human beta globin locus control region. NucleicAcids Res 19:1413–1419.

Ray R, Snyder RC, Thomas S, Koller CA, Miller DM (1989) Mithramycinblocks protein binding and function of the SV40 early promoter. J ClinInvest 83:2003–2007.

Robbins J (2001) KCNQ potassium channels: physiology, pathophysiology,and pharmacology. Pharmacol Ther 90:1–19.

Rong Y, Baudry M (1996) Seizure activity results in a rapid induction ofnuclear factor-kappa B in adult but not juvenile rat limbic structures.J Neurochem 67:662– 668.

Roza C, Lopez-Garcia JA (2008) Retigabine, the specific KCNQ channelopener, blocks ectopic discharges in axotomized sensory fibres. Pain138:537–545.

Schoenherr CJ, Anderson DJ (1995) The neuron-restrictive silencer factor(NRSF): a coordinate repressor of multiple neuron-specific genes. Science267:1360 –1363.

Schroeder BC, Hechenberger M, Weinreich F, Kubisch C, Jentsch TJ (2000)KCNQ5, a novel potassium channel broadly expressed in brain, mediatesM-type currents. J Biol Chem 275:24089 –24095.

Singh NA, Charlier C, Stauffer D, DuPont BR, Leach RJ, Melis R, Ronen GM,Bjerre I, Quattlebaum T, Murphy JV, McHarg ML, Gagnon D, RosalesTO, Peiffer A, Anderson VE, Leppert M (1998) A novel potassium chan-nel gene, KCNQ2, is mutated in an inherited epilepsy of newborns. NatGenet 18:25–29.

Sinha S, Fuchs E (2001) Identification and dissection of an enhancer con-trolling epithelial gene expression in skin. Proc Natl Acad Sci U S A98:2455–2460.

Spencer EM, Chandler KE, Haddley K, Howard MR, Hughes D, Belyaev ND,Coulson JM, Stewart JP, Buckley NJ, Kipar A, Walker MC, Quinn JP(2006) Regulation and role of REST and REST4 variants in modulationof gene expression in in vivo and in vitro in epilepsy models. NeurobiolDis 24:41–52.

Steiner C, Muller M, Baniahmad A, Renkawitz R (1987) Lysozyme geneactivity in chicken macrophages is controlled by positive and negativeregulatory elements. Nucleic Acids Res 15:4163– 4178.

Sun YM, Greenway DJ, Johnson R, Street M, Belyaev ND, Deuchars J, Bee T,Wilde S, Buckley NJ (2005) Distinct profiles of REST interactions withits target genes at different stages of neuronal development. Mol Biol Cell16:5630 –5638.

Surti TS, Jan LY (2005) A potassium channel, the M-channel, as a therapeu-tic target. Curr Opin Investig Drugs 6:704 –711.

Tinel N, Lauritzen I, Chouabe C, Lazdunski M, Borsotto M (1998) TheKCNQ2 potassium channel: splice variants, functional and developmen-

tal expression. Brain localization and comparison with KCNQ3. FEBSLett 438:171–176.

Uchida H, Ma L, Ueda H (2010) Epigenetic gene silencing underlies C-fiberdysfunctions in neuropathic pain. J Neurosci 30:4806 – 4814.

Wang HS, Pan Z, Shi W, Brown BS, Wymore RS, Cohen IS, Dixon JE, McKinnonD (1998) KCNQ2 and KCNQ3 potassium channel subunits: molecularcorrelates of the M-channel. Science 282:1890–1893.

Wickenden AD, McNaughton-Smith G (2009) Kv7 channels as targets forthe treatment of pain. Curr Pharm Des 15:1773–1798.

Wickenden AD, Krajewski JL, London B, Wagoner PK, Wilson WA, Clark S,Roeloffs R, McNaughton-Smith G, Rigdon GC (2008) N-(6-chloro-pyridin-3-yl)-3,4-difluoro-benzamide (ICA-27243): a novel, selectiveKCNQ2/Q3 potassium channel activator. Mol Pharmacol 73:977–986.

Wong J, Shi YB, Wolffe AP (1995) A role for nucleosome assembly in bothsilencing and activation of the Xenopus TR beta A gene by the thyroidhormone receptor. Genes Dev 9:2696 –2711.

Wong J, Shi YB, Wolffe AP (1997) Determinants of chromatin disruptionand transcriptional regulation instigated by the thyroid hormone recep-tor: hormone-regulated chromatin disruption is not sufficient for tran-scriptional activation. EMBO J 16:3158 –3171.

Wood IC, Roopra A, Buckley NJ (1996) Neural specific expression of the m4muscarinic acetylcholine receptor gene is mediated by a RE1/NRSE-typesilencing element. J Biol Chem 271:14221–14225.

Wood IC, Belyaev ND, Bruce AW, Jones C, Mistry M, Roopra A, Buckley NJ(2003) Interaction of the repressor element 1-silencing transcription fac-tor (REST) with target genes. J Mol Biol 334:863– 874.

Xiao HS, Huang QH, Zhang FX, Bao L, Lu YJ, Guo C, Yang L, Huang WJ, FuG, Xu SH, Cheng XP, Yan Q, Zhu ZD, Zhang X, Chen Z, Han ZG, ZhangX (2002) Identification of gene expression profile of dorsal root gan-glion in the rat peripheral axotomy model of neuropathic pain. Proc NatlAcad Sci U S A 99:8360 – 8365.

Xiao JF, Fischer C, Steinlein OK, JianFeng X (2001) Cloning and mutationanalysis of the human potassium channel KCNQ2 gene promoter. Neu-roreport 12:3733–3739.

Yang WP, Levesque PC, Little WA, Conder ML, Ramakrishnan P, NeubauerMG, Blanar MA (1998) Functional expression of two KvLQT1-relatedpotassium channels responsible for an inherited idiopathic epilepsy. J BiolChem 273:19419 –19423.

Zaika O, Lara LS, Gamper N, Hilgemann DW, Jaffe DB, Shapiro MS (2006)Angiotensin II regulates neuronal excitability via phosphatidylinositol4,5-bisphosphate-dependent modulation of Kv7 (M-type) K � channels.J Physiol 575:49 – 67.

Zuccato C, Tartari M, Crotti A, Goffredo D, Valenza M, Conti L, CataudellaT, Leavitt BR, Hayden MR, Timmusk T, Rigamonti D, Cattaneo E (2003)Huntingtin interacts with REST/NRSF to modulate the transcription ofNRSE-controlled neuronal genes. Nat Genet 35:76 – 83.

Zuccato C, Belyaev N, Conforti P, Ooi L, Tartari M, Papadimou E, MacDonaldM, Fossale E, Zeitlin S, Buckley N, Cattaneo E (2007) Widespread disrup-tion of repressor element-1 silencing transcription factor/neuron-restrictivesilencer factor occupancy at its target genes in Huntington’s disease. J Neu-rosci 27:6972–6983.

Mucha et al. • KCNQ Gene Regulation J. Neurosci., October 6, 2010 • 30(40):13235–13245 • 13245